Skip to main content
Kanaeru (Japanese for “to fulfill”) is an outcome delivery platform. Kanaeru doesn’t hand you a code assistant. You give it a vision, a description of what you want to exist. A fleet of coding agents then turns the vision into requirements, plans the work, builds it in isolated Git worktrees, reviews the results, and delivers working software as pull requests. Each cycle ends with a retrospective, and what it teaches carries into the next one. You work with Kana. She has read everything about your project and thought it through before you ask. She takes a position, explains the reasoning behind it, and shows you the alternatives she rejected. She keeps you in the loop at the points that matter: she shows you what she understood before any code is written, asks when something is yours to decide, and hands you evidence before you accept the result. Koa is Kana’s operator agent. Hand an outcome to him with Ask Koa to take over and he runs its whole lifecycle in Kana for you, from planning through delivery, one Koa per outcome. He decides from what you approved and the evidence in front of him, and stays in touch through your human inbox, Waiting on you, only when a call is genuinely yours. Choose Take over from Koa to take the wheel back at any time.
Kanaeru is in an invite-only closed beta, with no published pricing yet. To get an invite, join the waitlist on the Kanaeru Software Factory page. Your invite email has the link to sign in.

Quickstart

Sign in, connect a machine, and take your first vision to a pull request.

How Kanaeru works

Visions, outcomes, blueprints, tasks, and the lifecycle that connects them.

Connect a machine

Run agents on your own computer or on a cloud machine.

Kana CLI

Drive Kanaeru from your terminal.

​
See it in action

Follow Nadia as she takes one idea through Kanaeru, one session at a time.

1. Vision session

From a rough idea to a roadmap and a first outcome.

2. Planning session

From an outcome to tasks, journeys, and blueprints you can review.

3. Review session

How review caught a bug before any code was written.

​
What makes Kanaeru different

Before any code is written, Kana produces blueprints: mockups, user journeys, API contracts, data models, and architecture diagrams that show what she understood. Mismatches surface as understanding gaps you resolve up front, not as rework after delivery.
An outcome is one cycle of work that ends in something you can use. Kanaeru plans it into tasks, runs each task through a pipeline of implementation and checks, opens pull requests, responds to review and CI, and closes the cycle with a delivery review and a retrospective.
Each task gets its own Git worktree on an agent machine, so several tasks and several outcomes can run at once without stepping on each other. Machines run the coding agent you choose: Claude Code, Codex, Copilot CLI, and more.
A session isn’t tied to one computer, one coding agent, or one person. Move it between cloud and local machines in any direction, switch from Claude Code to Codex or any other runtime between messages, and let any teammate take the next turn or queue a message. Their turn runs on their own machine and sign-ins. The conversation, the code, and the plan come along each time. See Sessions that move.
Per kind of work, you decide whether Kana acts alone, acts and tells you, or waits for your approval. Spend caps and permissions sit with your organization’s owners and admins.
Every outcome ends with a retrospective on what went well, what didn’t, and what it cost. When the outcome completes, its findings are saved as learnings, and active learnings are given to later Kana sessions, so the next outcome starts from what the last one learned. You can review those learnings with kana memory. See Learning loop.

​
Ways to work with Kanaeru

Automations work across all of these: scheduled runs, API triggers, and signed webhooks that start Kana work without anyone opening a chat.